Preface: I am not an expert in wireless networking, nor Aruba products and capabilities, so if I write something incorrect, don’t send me to hell. Our company got a device like this and my colleague, Gabor Juhasz set it up, so he has the credit for all this.
Wireless network is no longer a luxury item and a modern office cannot work without it, but it is not uncommon to have them deployed to industrial areas and factories. From I experience I remember a client which used wireless connected barcode scanners to sort parcels and shipments, and this was six years ago. And believe me that is mission critical. Planned and currently built office spaces are often without any wired network which reduces the cost and also giving higher flexibility as no need to buy switches with at ports at least equal to the user count. So by now you are clear with the idea why it must be resilient and reliable.
And before we dive deep into this let’s clear two things when saying “working”:
- I can connect to it
- I can connect to it + able to reach the necessary applications and websites
Hungary is not the prime place for ITO, although we have many SSCs but typical Hungarian companies rarely outsource their IT as a whole. Regardless of this I am sure that the team responsible for wireless network is centralized and they are not distributed to all sites with wifi coverage. On a site without wireless expert, the troubleshooting is often cumbersome since even simple tasks, like if an AP is transmitting the SSID or not, which channel it is or why it cannot reach the wireless controller, can be daunting. If there are multiple access points on can be switched into spectrum analyzer mode but this is not like searching the root cause from the user perspective, instead this investigates from the infrastructure side.
Aruba User Experience Insight – aka Cape Networks
This is there Aruba User Experience Insight comes into play. Since HPE bought Cape Networks this is a renamed device. It is a sensor which is like a small access point.
One/multiple sensor(s) like this can deliver real time insight and active monitoring of the wireless network and the needed target systems by using it, in some cases the response time, the noise and retry rate too. If a sensor is deployed as static device like an AP it acts like a expert user, it sits behind his/her computer 7×24 and repeating the test we tell them to repeat continuously. I believe a sensor is cheaper than having a real flesh user with this role 😀
If we use it as a mobile station we can move them into the location where a possible issue is reported. Issues that are not there always, but intermittent. This way an investigation can boil down a specific root cause in a specific time frame.
How is it working?
Let me be clear. this is a cloud based solution. It utilizes Aruba’s cloud – NOT Aruba Cloud as that is for some reason using the same brand – to deliver the front-end and the necessary logic. No option to have this on premises.
The sensor itself is sleek, it does have only one PoE capable Gigabit Ethernet (10/100/1000)T and an external power connector. Question immediately arises, why wired network? The simple answer is it can run tests both on wireless network and wired network, furthermore if there is an issue with the wireless network, the device remains manageable through it’s wired connection. This is required since if it cannot reach the internet, alerts will not reach the Aruba cloud, so we will only see that the sensor became offline.
The cloud dashboard
After authentication the first page we will see is the main one. It is clean and clear, not hard to spot when an issue is affecting our environment.
The experience is about generally describing if everything is as it should be. The Services column is showing the tested state of the configured services. The Internal part is the actual state of our configured services and apps. The External is pretty similar to the Internal part, but showing the public services we define – like O365 or in my case the company’s public website 99999.hu.
If we click on anything, for example on Wi-Fi, than a window opens with more detailed information with the option to select the range of the view. Down below you can see the state of the wireless network based on 5 minute interval measurements. It tells which channel the AP used, the utilization of that channel and the time required to associate with that AP. The left bottom diagram shows the alerts and their appearance.
After clicking on this it will display in depth data.
Let’s get back to the main dashboard where now we select the 99999.hu under Services. These tests were based on checking ports 80/443, also if HTTP response is correct and can validate the validity of the SSL certs – if it expires this will tell you for sure. Can do ICMP tests and also measure the response time of HTTP request.
The first menu is where the sensor list can be seen and the groups they can be added. This helps with assigning tests quickly to sensors if we want to run them on multiple groups. Group can be anything, a site, a floor, a building.
The next ab is networks which is used to configure the wireless networks we want to monitor and also the wired network. Device is proxy capable so you need that to reach services, configure that here too.
The testing tab where it is defined which internal and external services we test and what kind of tests we want to run on them.
The “Thresholds” is the section were the tiny little bits can be customized. Like what we would like to do when an SSID disappears, when to alert. List is quite extensive I must say.
The “Alerts” tab is where we can define what should happen if something is triggered or under/above a given threshold. It can publish to an external target (Slack, ServiceNow or Iris), shall it send an email, a weekly mail, make a coffee etc.
Let’s create a test
Little selfish but also showcasing it with style, I’ll be defining a test for my own website, so newman.cloud. Down below this will be an external test and also custom – because under predefined only major players are enlisted. You can see there is an option to perform iperf based internal/external tests too if you want to put some load on the network and measure it’s performance from time to time. The Generic part is where the tested ports can be configured, in case of webserver the known ports are given.
I go for this.
After some minutes the first results are in and can see them.
Test from real life – real impact
While I was editing this post users have noticed an issue with the internet destined traffic of theirs. I checked what the system knows about this.
Dashboard was showing the same faces I saw on real users’ faces. Also that the wireless network, so as the DHCP service, the default gateway are all healthy, but DNS is showing some errors.
After selecting the DNS, a summary opens and the timeframe is drawn, so reveals when the issue started.
“Ongoing” tells that the issue is still persists. By clicking on this the detailed test times are their state is listed. Quite remarkable to see how many tests are there in the background, it traces the DNS servers, check the configured Cloudflare DNS servers.
The root cause was simply the unavailable DNS service, all other things were fine. After resolution it immediately reported that issue is solved and also changes the state from “Ongoing” to “Resolved”, but keeping the test results between issue start and end date.
Summary and conclusion
My conclusion is simple, this device/service is quite useful and can help continuous monitoring and quickly give a hand in troubleshooting is something is not working as expected. Of course you can live without it, but if the user experience is important for you or have multiple and would like to monitor the availability of your internal and public services this is something you should consider. One or two sensors like this with their subscription will quickly return the investment on Aruba Experience.